OpenAI’s EU Business AI Finding: Agentic AI And The Denominator Behind It

🔍 Read the full analysis: OpenAI’s EU Business AI Finding: Agentic AI And The Denominator Behind It on ThorstenMeyerAI.com

Buying for a business?Offer from Amazon

Get business pricing on monitors, keyboards and dev gear

• Business-only prices and quantity discounts
• Tax-exempt purchasing
• Multiple users, one account, clear invoices

Create a free Business account →Free to register

As an affiliate, we earn on qualifying purchases.Hide

OpenAI said Codex and ChatGPT Work produced 55% of output tokens in its ChatGPT Enterprise and Business accounts across the 27 EU member states by September 30, 2026, a finding that points to growing use of its AI agents but does not mean agents handled most users’ tasks or generated most business value. The company’s new report also says weekly active users of the tools at businesses with 200 or fewer users rose 8.1 times from the first week of May to the week ending October 5.

OpenAI’s Global Affairs team published The EU’s next chapter in AI is agentic on October 14, alongside a ten-page report from OpenAI Economic Research, Europe’s Agentic Opportunity: Moving from breadth to depth with AI agents. The figures cover ChatGPT Enterprise and ChatGPT Business accounts in the EU’s 27 member states. OpenAI defines the agentic tools counted as Codex, its coding agent, and ChatGPT Work, an agent mode launched in July 2026 that can work with files, browsers and company tools.

By September 30, Codex accounted for 34% of output tokens and ChatGPT Work for 21%, with ordinary ChatGPT accounting for the remaining 45%. The 55% is a share of text generated by the models, not a share of interactions, employees or companies using agents. OpenAI’s Enterprise Signals page describes tokens as an imperfect measure of business value and says agentic tasks can require substantially more computation.

In businesses with 200 or fewer users, OpenAI reports that weekly active users of Codex or ChatGPT Work increased 8.1 times between the first week of May and the week ending October 5. Every EU member state recorded growth of at least five times. Belgium had the highest reported growth at 12.1 times, followed by Italy at 10.8 and Luxembourg at 10.0. OpenAI gives no absolute user counts in the source material.

What the Token Share Captures

The report offers a window into how customers of one AI provider use its products, and suggests that agent tools are taking a larger role in work done through OpenAI’s business services. It does not establish that agents now dominate business activity across Europe. The distinction matters because an agent can generate many tokens while reading files, planning, producing work and checking results; a short chat exchange may generate far fewer.

OpenAI’s own caveat limits what readers can infer from the headline share. A 55% token share is evidence that agents account for a large portion of the company’s generated output in these accounts. It is not, by itself, evidence that 55% of AI tasks are agentic, that most employees use agents, or that the tools have produced a corresponding share of economic value. The report’s conclusion—that Europe should turn broad adoption into deeper use—is a policy argument from the company, not a result directly measured by the token figure.

The reported 8.1-times growth also needs its starting point. It compares weekly active users with a week in early May, before ChatGPT Work launched in July. OpenAI says a major part of the increase followed that launch. The figures signal rapid uptake of the products, but without absolute user totals or a longer time series, they do not show how widespread use is across Europe’s small businesses or whether the pace will continue.

How Europe Compares in OpenAI’s Data

The report also compares output among companies it calls frontier firms, defined as the top 10% by output tokens per active user. In September, EU frontier firms produced about four times as many tokens per user as typical EU firms. OpenAI says frontier firms outside the EU produced more than twice as many tokens per user as the EU frontier. EU companies accounted for about 10% of frontier firms worldwide, compared with about 65% in North America.

OpenAI reports wide differences among member states in agentic share: Lithuania was at 85% and Cyprus at 81%, while Slovenia was at 22% and Austria at 26%. Germany’s share was 51%, ranking 17th of 27. For small-business weekly active users, growth ranged from at least fivefold in every country to 12.1 times in Belgium. These are figures from OpenAI’s own customer accounts, not a census of all businesses in each country.

The source report frames Europe’s next step as moving from breadth of adoption to deeper use, including worker skills and confidence and access to advanced capabilities. Its comparisons describe activity on OpenAI’s services; they do not independently measure national AI readiness or compare the full market across providers.

What OpenAI’s Figures Leave Out

The report does not provide absolute account or user counts for the small-business growth figure, nor sample sizes for each member state in the country comparisons. Without those details, readers cannot assess how representative the data is or how much a small number of large customers may affect a country’s share. The report’s figures describe OpenAI customers, not all EU businesses.

It is also unclear how much of the measured increase reflects sustained changes in work practices rather than the launch and early adoption of ChatGPT Work. The source says the launch contributed substantially to growth, but does not quantify its precise contribution to small-business user increases. Nor do the supplied figures establish whether agent use improves productivity, revenue or other business outcomes.

The token share is not a task count. An illustrative calculation in the source article assumes an agent run generates 10 to 30 times the output of a chat reply; that assumption is not a finding in OpenAI’s report. The report does not provide the interaction-level data needed to convert its token share into the proportion of agentic tasks.

Evidence Needed Beyond Adoption

OpenAI’s report points to worker skills, confidence and access to advanced tools as areas for Europe to address. The next useful evidence would include absolute adoption figures, country-level sample sizes and longer-term usage data, as well as measures of outcomes such as time saved or work completed. Those details would help distinguish a rapid rise in product activity from broad, lasting changes in business practice.

For now, the report establishes what OpenAI says is happening inside its own EU business accounts through late September and early October 2026. Whether the patterns extend across other AI providers, sectors and businesses—and whether higher agent use translates into measurable economic gains—remains unresolved.

Key Questions

What does OpenAI mean by 55% of output being agentic?

It means Codex and ChatGPT Work generated 55% of the output tokens in the covered EU ChatGPT Enterprise and Business accounts by September 30, 2026. It does not mean 55% of users, tasks or business value involved agents.

Which tools did OpenAI count as agents?

The report counts Codex, OpenAI’s coding agent, and ChatGPT Work, an agent mode launched in July 2026 that can work with files, browsers and company tools.

How fast did small-business use grow?

OpenAI reports that weekly active users of the two tools at businesses with 200 or fewer users rose 8.1 times from the first week of May to the week ending October 5. The source gives no absolute user totals.

Does the report show that Europe’s businesses are behind?

OpenAI says its EU frontier firms generated less than half as many tokens per user as frontier firms outside the EU. That comparison is based on OpenAI customer output and does not measure the full European AI market or prove a broad economic gap.

Source: ThorstenMeyerAI.com